Researching the Institutional Structure of Technological Innovation: Working with IP Data - Wednesday Workshop The Structure of our Session Motivation: Controversies Overview of IP Data: Sources and Uses Questions 5 minutes 20 minutes 15 minutes ESNIE 2006 Cargèse, Corsica 17 May 2006 s in Researching the Institutional Structure of Innovation Questions Discussions - Questions 20 minutes 15 minutes 15 minutes Remember this drawing? A Broken Patent System? astly, it should be noted that because pulling alternately on one chain and then the other resembles in some measure the movements one would use to swing from vines in a dense jungle forest, the swinging method of the present invention may be referred to by the present inventor and his sister as "Tarzan" swinging. The user may even choose to produce a Tarzan-type yell while swinging in the manner described, which more accurately replicates swinging on vines in a dense jungle forest. Actual jungle forestry is not required. icenses are available from the inventor upon request. Jaffe & erner: Innovation and its Discontents 350,000 Figure 1 USPTO Utility Patents 1965-2002 Patent Regime reconfigured 1982: Founding of the CAFC Early 90s: USPTO Fee structure changed Resulting parade of horribles Patents on inventions that are trivially obvious Strong patents are potent strategic weapons, e.g. semiconductors Huge court awards tax competitors Knowledge is slow to reach examiners, esp. in new technologies Number 300,000 250,000 200,000 150,000 100,000 50,000 0 1965 1970 1975 1980 1985 1990 1995 2000 Year Patent applications Patent grants lag two years
Why patents? Economic Rationale Societal costs (Welfare oss) of Patent System Deadweight loss from patent monopoly Transaction costs Greater benefits to society (Welfare Gains) Create incentives for invention Require disclosure of invention to public Knowledge spillovers may enable follow-on innovation Create incentives for commercialization of invention by inventor, or others Support markets for Intellectual Property May enable vertical specialization in industries Forces eroding Welfare Gains ow Quality acking requisite novelty, non-obviousness, utility Uncertainty Over final boundaries of the disclosure Over the validity of the property right Under- or misdirected investments by inventor in the patented technology by competitors in competing technologies Adds transaction costs to commercialization, technology transfer (licensing), developing markets for IP Conferring market power to trivial innovations with little or no social welfare Create an environment inviting to costly litigation Posture: Policy demand for Reform Academy Science, Technology and Economic Policy Board (STEP) of the National Academies Committee on IPRs October 2002 meetings US Government Agencies USPTO Strategic Plan for the 21 st Century (2002) FTC To Promote Innovation: The Proper Balance of Competition and Patent aw and Policy (2003) Interest groups AIPA (2004), IBM, other large firms, BSA US Congress HR 2795 (2005); Rep s Berman (D-Ca) &. Smith (R-Tx) IP-data based research: current knowledge My main thrust... Questions for this part of the presentation What types of IP information are available? How have these data been used to study the institutions of innovation? Where may I obtain IP data? What are some opportunities, and challenges, in using IP data? Main Species of IP Protection: What s in them, and how can it help me as a researcher? Observable information: what does it look like? Patent Government-granted right to prevent others from using a stated technology Administrative process, with substantial applicant input imited in time, scope monopoly? Sort of, but not *really*... Copyright Government-granted, and natural right protecting authors, creators imited? in time, limited to expression Trademark Government-granted protection for brand marks, consumer protection imited in subject matter, unlimited in time but actively protect, maintain Secret Protection? In the US: valuable information, reasonable efforts, unlimited in time but independent discovery Observable?
Observable information: IP data Observable information: the territories First, some things that are not in the data What s not observable? No information on products No performance data per se No sales data United States: US PTO Europe: EPC-EPO Japan: JPO Why not? Purpose of disclosure requirements: dissemination of technical and legal knowledge, not necessary economically relevant knowledge What s observable then? A bunch of information on applicants, inventors, legal terms, technology What information exactly? Depends on geography. I will focus primarily on European and US patent data Observable information: the EP granting procedure Observable information: the EP granting procedure Formal Examination Publication of the and Search Report Applicant and Inventor Information Backward References Filing Path (Strategy) Granting of the Version Approved by the Applicant Substantial Examination Yes Request for examination Grant Final claims Acceleration PCT I / II Opposition Appeal National Rights No Withdrawal/Rejection of the Patent Opposition & Appeal & Designated States Granting Procedure Outcome Observable information: US granting procedure Observable information: US granting procedure Granting of the Version Approved by the Applicant Examination Publication (since 2001) of the (but caveat) Negotiation, claim amendment No Grant Backward References (final) Final claims Reexamination & Applicant and Inventor Information Backward References (initial) Initial Claims Filing Path (Strategy) Reexamination itigation Withdrawal/Rejection of the Patent itigation & Granting Process
Observable information: a list of data Relevant information for management scholars Immediately observable Applicant data Inventor data Technology classes Backward references Familiy size Filing routes (patenting strategy ) itigation Computable Forward citations (time lag) Originality Generality (time lag) Interpretating information Equivalents or Family" Information Contains: Designated countries (incl. country code) Has been applied to the study of: Innovation Valuation of IP Institutional Economics Some references: anjouw/pakes/putnam (1996); Graham/Hall/Harhoff/Mowery (2003); Graham/Harhoff (2006); Jensen/Palangkaraya/Webster (2006). Interpretating information Timing Dates, Continuations, PCT Contains: Information on the timing of initial spark, strategic choices during application process Has been applied to the study of: Valuation of IP Competition analysis Firm Strategy Some references: Quillen/Webster (2003); Graham/Mowery (2004); Reitzig (2004); Graham/Harhoff (2006); Jensen/Palangkaraya/Webster (2006). Interpretating information Post-grant Reviews (e.g., litigation, opposition, etc.) Contains: Information on type of suit challenges (validity or infringement), names of parties, outcomes Has been applied to the study of: Valuation of IP Competition Analysis Institutional Economics Some references: erner (1995); anjouw/schankerman (2001); Harhoff/Scherer/Vopel (2003); Somaya (2003); Graham/Hall/Harhoff/Mowery (2003); Harhoff/Reizig (2004); anjouw/erner (2004); Graham/Harhoff (2006) Interpretating information Forward Citations Contains: Information on how often/by whom a patent is referenced as relevant prior art during subsequent patent examination processes Has been applied to the study of: Valuation of IP Social network theory Technological pathways Knowledge flows Some references: Trajtenberg (1990); Trajtenberg/Jaffe/Henderson (1997); Stuart (1998); Mowery/A.Ziedonis (2001); Alcacer/Gittleman (2003); Hall/Jaffe/Trajtenberg (2005) Interpreting information Research areas previously explored using patent data: Competition analysis Firm Strategy Innovation Institutional Economics Knowledge flows Social network theory Strategic alliances Valuation of intangible resources And in the future?...
Collecting information Data sources Free but raw, offical registers, designed for case-based research. No computed indicators EPOINE USPTO Free, not raw, but used NBER (Hall, Jaffe, and Trajtenberg, 2000) US patent data complete 1975-1999 (with some data 1963-2002) Standard indicators (incl. forward cites) computed NSF (Cockburn, et al., 2005) Multi-year update of the NBER data, with link-outs Commercial providers Micropatent Derwent Using information Opportunities and Challenges?? I: Post-Grant Reviews The costs to using data have (and continue to) fall NBER/NSF database; Computing power Uncertain rewards to using the current data Does the use of data become expected, or passé? ikely rewards at the frontier New applications, employing greater institutional knowledge (and methodological skills) Other Data Copyright Trademark Secrecy? Economic criticisms: US itigation Costly: Estimates $0.5-4M per suit; $500K per claim some as high as $20M in biotech, $48M for Polaroid. Prolonged: Estimated 31 months for trial Burdensome: May not serve quality concerns Patent afforded a presumption of validity; born valid Generally, patentee holds the litigation trigger Burden of proof set high: clear and convincing standard Judge and/or jury may have limited expertise Escalation: Increasing rates, and sheer numbers Estimates changed little from 1970s to early 1990s: 1-2% Rate rose in late 1990s; current research 1998-2000: 3% As number of issued patents explodes => litigation explodes Administrative Alternatives to itigation: US, EU (EPC) United States patent challenges Reexamination post-issue (during the life of the patent) Ex parte, inter partes EU (EPO) patent challenges Opposition post-issue (within 9 mos.) itigation for validity or infringement in national courts
EPO Opposition: Distinctions Tab. 4: Results - EP applications Administrative, non-judicial process Examiners (technical experts) hear challenge Administrative judges on appeal Adversarial proceeding initiated by any third party de novo hearing Validity questions (not infringement): Patent may be challenged on grounds of patentability Subject matter, inadequate disclosure, scope of grant larger than original application Much lower cost than litigation, cost ( 20-30K) In operation, 8% of patents are opposed 1/3 of these are revoked outright, with a further 1/3 partially revoked. But Time limit - 9 months of patent grant But are generated slowly (average 1.9 years) Tab. 5: Results - Oppositions Fig. 5: Opposition Frequency Welfare Analysis Cost-benefit analysis We use evidence to inform us: costs, rates itigation rates 1.1% - 3.2% Opposition rates equivalents of non-litigated US patents 6% equivalents of litigated US patents 20% : revocation about 1/3, amendment about 1/3 : benefits avoided litigation costs $4m (AIPA) reduced (erroneously granted) market power ($1-$4m) Appeal rates 52% for equivalents of litigated US patents 32.5% for equivalents of non-litigated US patents Welfare Calculations (1.1) Saved litigation expenses W = p P p ) S 1 O, ( pr, + 0.5 ppr, (1.2) Removing excess market power W = ) S 2 ( 1 p) P po, N ( pr, N + 0.5 pprn, (1.3) Costs of post-grant review C = p P p O, ( C O + ( p A, C A )) + 1 p ) P p ( C + ( p ( O, N O A, N C A N )
Tab: 8: Estimates Conclusions US litigation system is costly and burdensome to challengers those with incentives, information EPO examination does not disproportionately exclude patents whose US equivalents were litigated. Opposition rates for equivalents to litigated patents substantially higher We find that the welfare gains from an opposition system may be substantial The main source of these welfare gains is not foregone litigation, but elimination of excess market power. Caveat: Any benefits will likely be eroded with increasing cost of opposition. II: US Continuation Patents US Patent Continuation practice 2 Truisms The patent office is unable to ever finally refuse to grant a patent The patent office is unable to ever finally grant a patent The culprit? The continuation application Rambus, Incorporated: An Intellectual Property Company This application is a divisional of application Ser. No. 08/710574, filed Sep. 19, 1996, now abandoned, which is a continuation of application Ser. No. 08/469,490 filed Jun. 6, 1995, now abandoned, which is a continuation of application Ser. No. 07/847,961 filed Mar. 5, 1992, now abandoned, which is a divisional of application Ser. No. 07/510,898 filed Apr. 18, 1990 now abandoned.
Institutional variation Continuation, Division, and Continuation-in- Part procedures FIGURE : Schematic of the Continuation Patenting Process time Continuation i t 0 t 1 t 2 patent Patent Grant Division i i abandon abandon patent patent Inventor Submit Forego Examiner patentable not patentable Inventor Inventor Continuation C -in-part i new matter patent negotiation, feedback Terminate abandon 90% Figure 2.10: Continuation Shares, Top Biotechnology Firm s' Patents 1985-1999 (Amgen, Chiron, Genentech, Immunex) Figure 4: Continuation shares of issued patents in selected sectors, 1975-2000 (chemicals, computers, electronics devices, combustion engines) % Patents with Continuation 80% 70% 60% 50% 40% 30% 20% 10% 0% 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% chemical computer electronic devices combustion engine Chemical Computer Electronic Engine 1985 1987 1989 1991 1993 1995 1997 1999 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 Supplemental: U.S. Regime Changes: 1995, 1999 Continuation priority was arguably more valuable pre-1995 Patent term 17 years measured from date of issue Uruguay Round GATT, effective June 1995 Int l harmonization; reduce incentives for continuation Patent term now 20 years from date of application Filings after June 95, maximum 20 years to both prosecute and protect an invention Forces a choice, ex ante secrecy versus ex post protection 1999 Inventors Protection Act Publication of most applications after 18 months Excepted: special class of US-only inventions Conclusions Opportunities abound for productive research in this area Changing research environment While the costs to using these data have fallen dramatically the rewards to understanding, and being able to productively analyze, the institutional details has increased